Are Dog Bite Statistics Reliable?

Yesterday, I wrote about Dr. Mark Wulkan, who says that some breeds of dogs have no place around kids. Not surprisingly, he singles out the usual suspects - pit bulls and Rottweilers.

We've poked holes in his conclusion. Now let's take a look at his statistics. When it comes to tracking dog bites, numbers are tricky things.

First, not all dog bites are the same. A major source of data for dog bite studies, like the one Wulkan cites, are incidents reported by animal control or public health agencies. According to the National Canine Research Council, an incident could be anything from an accidental and playful scratch or nip to a full-blown attack by a poorly-trained, aggressive dog. Those reports also include incidents where a scared or injured animal bites a rescuer or vet.

Now, most bite reports collected by animal control agencies do track the circumstances surrounding the bite, but they're usually reported to the public (and the scientific community) as just a raw number, not broken down by the attending circumstances. Most reporters and scientists just don't dig into the more subjective territory of circumstances, such as whether the dog was spayed or neutered, injured, chained, or left unsupervised with children — all non-breed specific factors that can lead to dog bites.

You also have to look at how reliable the breed-specific part of the data actually is. For example, my dog, a black pug, has been confused with everything from a mastiff puppy to a Boston Terrier to a bulldog. Looking at more factual evidence, the National Canine Research Council has an interesting piece(in .pdf format) showing pretty significant discrepancies between what people think a dog is, and what its DNA actually tells us he is. Add to that the extra confusion caused by mixed-breed dogs, and you can see why a visual report can be unreliable. Just because it looks like a German Shepherd doesn't mean it is a German Shepherd.

It's also worth nothing that some breeds of dogs might be victims of their own bad PR. If a dog bites or tries to bite someone, it's easy to assume that the dog is a pit bull, Rottweiler, or Doberman (to name a few breeds with undeservedly bad reputations) simply because those are the breeds that we hear about most often in terms of aggressive behavior. It's a vicious cycle, no pun intended.

All in all, Wulkan is looking at some decidedly questionable data. In fact, the sketchiness of the data is exactly why the Centers for Disease Control stopped tabulating dog bites by breed over ten years ago. In fact, CDC opposes breed-specific legislation for the very simple reason that you can't back up those policies with reliable statistics.

The CDC says that studies that indict particular breeds are unreliable. Dr. Randall Lockwood of the ASPCA calls the focus on specific breeds, "a 19th century epidemiological mindset." That's good enough for me. It ought to be good enough for Dr. Wulkan.

Photo credit: mutts and such